Optimal alluvial channel width under a bank stability constraint

To properly predict alluvial channel width using rational regime models, an analysis of bank stability must be included in the model. When bank stability is not considered, optimizations assuming maximum sediment transport capacity (MTC) typically under-predict alluvial channel width for natural and laboratory streams. Such discrepancies between regime model predictions and observed channel widths have been used to argue that optimizations such as MTC do not describe the behaviour of alluvial systems. However, rational regime models that explicitly consider bank stability exhibit no such bias and can predict alluvial channel widths quite accurately. We present an analysis of both laboratory and natural alluvial channels, using both kinds of models, and demonstrate the importance of bank stability in constraining optimization solutions. We also identify a scale effect, whereby the effect of vegetation on bank strength declines as the absolute scale of the system increases. We argue that comparisons of alluvial channel widths against predictions from rational regime models unconstrained by bank stability are inappropriate, because they introduce a known and quantifiable bias (toward under-prediction by the model) due to the absence of a bank stability constraint.